47

AUTUMN HAWK MIGRATION IN ONTARIO

STUDIED WITH RADAR

W. John RichardsonLGL Ltd., environmental research associates

44 Eglinton Ave. W., Toronto, Ontario M4R lAl, Canada

INTRODUCTION

The tendency of migrating hawks to concentrate at certain locationsis one of the features that makes study of their migration especiallyinteresting. This tendency to concentrate, coupled with their relativelylarge size and daytime migration, has made them one of the best knowngroups of birds insofar as migratory habits are concerned. Visual obser­vations at the known concentration points have provided quite detailedinformation about flight behavior and the timing of migration for eachspecies. This type of information is much less readily obtained forspecies of birds that do not concentrate and for those that migrate atnight.

The number of locations at which major concentrations of migratinghawks have been recorded systematically is quite small. The known concen­tration points are widely scattered. There are vast intermediate areasover which hawks certainly do migrate to some degree. Thus current know­ledge of hawk migration is based on systematic data from only a very fewrather unusual locations, scattered observations from other locations,and virtually no data from everywhere else.

There is, for example, little reliable information about how hawksget to the known concentration points, or about where they go from thosepoints. There is rarely more than speculation about what the hawks aredoing on days when they are not migrating in significant numbers over theusual concentration points. Are they migrating elsewhere, or are theygrounded awaiting more favorable weather? It is difficult to interpretthe data from concentration points in the absence of data from elsewhere.There is, for example, no way of determining from a single observationpoint whether the weather conditions associated with major flights at aconcentration point are truly those that induce maximal numbers of birdsto fly. They might instead be the conditions that produce maximal con­centration over the observation point. Clearly a broader perspective isrequired in order to examine these questions.

One way of obtaining this broader perspective is that used in theNew England Hawk Watch, described in this volume (page 137). This approachcan provide very detailed data, but it requires too much manpower to beapplied daily throughout a migration season.

The techniques of radar surveillance provide another way of obtain­ing broad-scale data on bird migration. Radars can provide informationabout the distribution, numbers, flight paths, speeds, and altitudes ofmigrants. Eastwood (1967) has provided a good introduction to radartechniques applicable to ornithology, and a review of results to that date.

Radars can detect birds regardless of visibility; hence fog, clouds,darkness, and rain need not preclude observations. On all but the lowest­powered radars, even the highest flying hawks are readily detected. Whileradars and associated equipment occasionally do break down~ they are

p. 47-58/n: M. Harwood (ed., 1975), Proc. NorthAm. Hawk Migration Cont., Syracuse, NY, April1974. Hawk Migr. Assoc. N. Amer. 165 p.

basically tireless.obtained radar dataseason.

By using time-lapse filming techniques, we have oftenfor well over 90% of the hours in an entire migration

'!'he lowest-powered radars, which include many marine, airborne andanti-aircraft fire-control radars, are capable of detecting individualhawks out to distances of at least a mile, and sometimes to several miles.

Medi~powered radars, such as are used for air traffic controlaround airports and for weather surveillance, can in theory detect indivi­dual hawks out to distances of about ten miles. In practice this probablyrarely occurS, partly because of various technical limitations associatedwith the difficulty of separating weak echoes from hawks and strong echoesfrom objects such as trees, buildings, and hillS, and partly because theseradars usually cannot resolve birds flying within a few hundred feet of oneanother. However, we have found that medi~powered radars of these typesdo detect groups of hawks at distances up to 20, 30, or even 40 miles fromthe radar site.

High-powered radars are used for en route air traffic control in theUnited States, for air surveillance by the U.S. and Canadian armed forces,and for missile tracking by the National Aeronautics and Space Administra­tion. These installations undoubtedly can detect hawks at even greaterranges, although to our knowledge they have not been used for this purpose.

While radars have these various advantages over visual observers,they have numerous disadvantages as well. Their resolution is sufficientlylow that the birds within some minimum volume of airspace cannot be dis­tinguished. '!'hus one cannot determine the number of birds in any givengroup with most radars. In general, the higher the radar power and thegreater the area surveyed, the poorer the resolution. Furthermore it isvery difficult to identify the birds appearing on radar screens. Whilebehavioral differences usually allow one to distinguish major groups suchas haWks, waterfowl, shorebirds and passerines, one can rarely make finerdistinctions. Finally, radars cannot detect low flying birds at rangesof many miles since these birds are below the radar horizon. I have dis­cussed these and other limitations in detail elsewhere (Richardson, 1972),

Clearly, the radar and visual techniques are complementary. Radarcan provide broad-scale and continuous data about flight directions, speeds,altitudes, and concentration areas, and more crudely about numbers aloft.Visual observers can provide detailed but very local information about thespecies involved, flock sizes, and small-scale flight behavior. Far morecan be learned by using both techniques together than by using either alone.

Hawks have been detected with radars by various observers in bothEurope and North America (Gehring, 1963: Houghton, 1970, 1974: Alerstam andUlfstrand, 1972: Evans and Lathbury, 1973). However, the present study inOntario and the concurrent study at Gibraltar are the first systematicanalyses of hawk migration that have employed radar data.

My dbjectives in this report are two-fold: to describe radar tech­niques useful in the study of hawk migration, and to present a preliminarysummary of our study of autumn hawk migration in Ontario. '!'he latter aspectwill be published in detail elsewhere.

48

RADARS USEFUL IN STUDIES OF HAWK MIGRATION

Three general types of radars can be used profitably in studies ofhawk movements: broad-beam surveillance radars, tracking radars, andspecial-purpose radars. We have used only surveillance radars for study­ing hawks, although we have used tracking and various special-purposeradars in other studies.

Broad-beam surveillance radars

Broad-beam surveillance radars are the most familiar types. Medium­and high-powered versions are used for air surveillance and air trafficcontrol. Lower-powered and less sophisticated versions are used on shipsfor navigation, but can be adapted for ornithological purposes (e.g.,Williams et al. 1972a). All of these radars involve a beam that is broadin vertical extent and narrow horizontally. This beam is rotated continu­oU$ly. When an ec~o is received, the distance of the target is determinedfrom the time taken for the echo to return. The direction is determined bythe direction in which the antenna was pointed when the echo was received.Because of the broad vertical extent of the beam, little or no informationabout altitude can be obtained from this type of radar. The distance anddirections of detected targets are displayed on the familiar circular radardisplay, called a Plan position Indicator or PPI, which is simply a map ofthe area. The more sophisticated surveillance radars have Moving TargetIndicator circuitry that suppresses all stationary echoes. This permitsone to detect aircraft and birds even when they are over areas from whichground echoes are being received.

By observing the PPI at one point in time, one can detect the pres­ence, locations, number, and concentration areas of echoes such as birds.By observing the PPI over a period of time, one can also detect thedirections and speeds of motion of the various echoes.

When one is observing an area 30 or more miles in radius, echoes frombirds take a long time to move across the PPI. Their motion during the fewseconds from one rotation of the beam to the next is almost imperceptible.However, time-lapse filming can provide an excellent record of this motion.The time-lapse technique has the additional advantage that one can reexaminethe record for a given period of time as often as desired. Furthermore acontinuous record can be obtained by the mere effort of changing the filmevery one to four days, depending upon filming rates and camera capacity.In most situations time-lapse techniques provide much more detailed in­formation than can be obtained directly or with a still camera.

We normally use l6mm equipment operating at about three frames perminute; thus a 200-foot roll of film, which costs about $20 to buy andprocess, will provide a continuous record over a 4B-hour period. Solman(1969) has described the filming technique. Williams and Mix (1973) havedeveloped a less expensive system using an Bmm camera.

Hawks migrating across southern Ontario in autumn and over Gibraltartypically appear on radar in a linear pattern readily distinguishable frommost other migrants (e. g., Figure 1). Simultaneous visual and radar obser­vations on numerous days around Toronto and at Hawk Cliff on Lake Erie haveshown that the vast majority of occasions with such patterns represent hawkflights. Other birds occasionally move in concentrated streams along lake­shores, but these flights usually are not accompanied by simultaneous linesof echoes inland. Furthermore only hawk flights begin at mid-morning; Blue

49

- ~ -- - --------~---

Jays, crows, swallows, and other potentially confusing species take off atsunrise. Thus I am confident that in southern Ontario, hawks are the birdsinvolved in any movement that first appears on the radar display at mid­morning and includes lines of echoes moving southwest inland. Many of theother isolated echoes moving southwest at the same time as the lines ofechoes are presumably hawks. However, we have no way of identifying them,and hence base our analyses on the lines alone.

Data can be abstracted from the time-lapse films using a variety oftechniques. Flight directions of individual echoes and the alignment ofthe lines of echoes can be estimated by eye. Alternatively, directions andspeeds can be measured by first tracing them onto a sheet of paper used asa screen, or by various photographic techniques. The positions of concen­trations of hawks can be determined by projecting the film onto a map atthe appropriate scale, or by recording their distance and direction fromthe radar site and transferring these coordinates to a map. The number ofhawk echoes over the whole display or in specific areas can be estimatedon an ordinal scale, or in some cases, by counting the numbers of indivi­dual echoes crossing some line perpendicular to the mean flight direction.The latter approach is workable only when the echoes in the lines are dis­tinct. In any event only the number of 'kettles' or flocks, not the numberof individuals, can be measured. Thus these radars can provide only a crudemeasurement of the relative numbers aloft at different places and times.Supplementary visual observations of flock sizes are very useful; the radarcan show the number of flocks over a broad area, while the visual observa­tions reveal the sizes of a sample of the flocks.

Medium- and high-powered surveillance radars (peak power 400 kilowattsto several megawatts) cost hundreds of thousands or millions of dollars,require frequent maintenance, and are rarely portable. Thus one is re­stricted to parasitizing preexisting installations. All major airports andsome smaller ones have medi~powered radars. High-powered military sur­veillance radars are scattered at intervals of about 250 miles acrosssouthern Canada, around the perimeter of the United States, and at variousother locations. In the U.S. there are additional radars of this typescattered across the country for en route air traffic control. Access toany of these radars usually can be obtained, although it often requiresseveral months of intermittent negotiation.

Low-powered marine surveillance radars are readily obtainable, new,at a cost of $5,000 to $30,000. These radars have peak power output ofthree to 50 kilowatts, and require very little maintenance. They are port­able in a van, and can be run by a small gasoline-powered generator or fromnormal power outlets. Their range is at most a few miles, but within thatrange they have higher resolution than do the more powerful surveillanceradars. Thus marine radars are well-suited for obtaining more accuratecounts of the number of birds passing by within two or three miles of anylocation to which one could drive a small vehicle. Marine radars generallydo not have circuitry for suppressing ground echoes, and hence provisionmust be made to avoid detecting objects at low elevation. This limits oneto surveying moderate or high altitudes.

A useful approach would be to use a moderate- or high-powered radarto survey hawk movements over a general region, a low-powered high-resolu­tion marine radar to obtain more detailed information near concentrationpoints within the coverage area of the more powerful radar, and visualobservers to obtain still more localized but more detailed data. Nosystematic study of this nature has been performed to date.

50

Tracking Radars

The second type of radar potentially of great value in studies ofhawk migration is the tracking radar. These radars have a narrow beam.Once the beam is aimed at a target of interest, it automatically remainsdirected at the target. The direction of the beam plus the time requiredfor the echo to return provide a precise measurement of the location ofthe target in three dimensions. The changes in position over time can beplotted automatically onto graphs in real time or recorded manually everyfew seconds. Alternatively or additionally, the positions can be recordedonto analogue or digital magnetic tape for future manual or computer anal­ysis. In addition, the strength of the echo received from each pulse ofenergy can be recorded. The variations in the echo intensity from a tar­get are called its signature. From the signature can be extracted infor­mation about size, wing-beat rate, and perhaps the attitude or orientationof the body relative to the radar beam. Bruderer and Steidinger (1972)and Williams et al. (1972b) describe tracking-radar techniques in moredetail. Thes;-techniques are now being applied in the study of hawkmigration at Gibraltar (Houghton, 1974).

A high-resolution tracking radar can provide very precise infor­mation about flight behavior. For example, the radius of turn, rate ofclimb and maximum altitude attained in a thermal could be measured easily.These parameters could be compared for different species, different weatherconditions, etc. Winds at any altitude of interest aloft can be measuredprecisely by using the same radar to track an ascending balloon.

Tracking radars are not well suited for general surveillance.Their beams are too narrow to examine more than a small fraction of thesky in anyone rotation, and they are not designed for continuous or un­attended operation. However, simultaneous observations with a surveillanceand tracking radar, such as are now being obtained at Gibraltar, canprovide excellent data.

For the purpose of studying hawk migration, a portable, low-powered,high-resolution tracking radar is probably preferable to a fixed, high­powered, moderate-resolution installation. Even the least expensivetracking radars are expensive. Thus the only practical alternatives areto obtain access to a preexisting fixed-base radar, to obtain a mobileradar on loan from NASA or the military, or to buy an obsolete militarysurplus radar. Each of these three approaches has been used successfullyby ornithologists. However, the third alternative, buying a surplus radar,is probably inadvisable because of the difficulty in maintaining old butcomplex electronic equipment for which parts are not readily obtainable.

Other types of radar

The third category of radars includes the wide variety of surveil­lance equipment with something other than vertically-broad beams rotatedhorizontally. Radars with vertically-narrow beams can provide informationabout the heights of migrants. However, with some of these radars it iseven more difficult than usual to identify birds. The usefulness of anyavailable radar must be evaluated on the basis of its individual charac­teristics and the data required.

One widely available class of radars, weather radars, provides mostof the advantages of surveillance radars plus a limited capability for

51-- ------------

height measurements and for determining the size of flocks (Able, 1970;Gauthreaux, 1970).

Military radar sites usually have 'nodding' height-finding radarsas well as broad-beam surveillance radars. Eastwood (1967: 197) andNisbet (1963) have described these height-finders. They can measure thealtitudes of birds over broad areas with a maximum possible error of about500 feet. However, they are often of little use in hilly terrain, whereechoes from the ground tend to obscure birds.

AUTUMN HAWK FLIGHTS OVER SOUTHERN ONTARIO

Results of visual observations

Previous visual observations have revealed massive southwestwardand westward flights near and along the north shores of Lakes Ontario andErie. The best-known observation point is Hawk Cliff, on the north shoreof Lake Erie. It is not uncommon to see 10,000 hawks moving west pastone point during a single september day, and as many as 70,000 have beenreported in one day at Hawk Cliff. The most abundant species is theBroad-winged Hawk, Buteo platypterus. It is widely known amongst localornithologists that the largest flights are to be expected along the lakeshores with northerly winds, cool temperatures, and rising pressure afterthe passage of a cold front, as a low-pressure area moves away to the eastand a high approaches from the west. Gunn (1954) and Haugh (1972) havedescribed autumn migration of hawks across southern Ontario.

West of Hawk Cliff the buteos tend to spread out inland ratherthan to change course to the west-southwest and follow the shore towardsPoint Pelee, although Sharp-shinned Hawks and perhaps others do tend toremain with the shore. Similar behavior has been suspected at Toronto.While hawks are frequently concentrated near and at the shoreline east ofand over the city, visual observations suggest that many then move inlandrather than follow the shore to the southwest.

On days when the winds have a southerly component, the main flightline along the north shore of Lake Erie is apparently a few miles inland.Haugh (1972) has concluded that varying positions of the main flight linebased on wind conditions are a major cause of year-to-year fluctuationsin the number of hawks seen. He reported that this was not the case atDerby Hill in spring, where the main flight line was quite distinct andonly a mile or so inland with onshore northerly winds.

Unfortunately, the visual data from inland southern Ontario aretoo sparse to be very useful for comparison with the shoreline data.However, flights of several hundred or even several thousand birds perday have been reported from a location about 20 miles north of the Torontolakeshore.

The radar view

Radar data from London, Ontario, including the Hawk Cliff area,were obtained through the autumns of 1965, 1966, and 1967. Similar datafrom a radar at Toronto were obtained in the autumns of 1964, 1965, 1967,1970, and 1973. Both radars were AASR-l broad-beam Air Traffic Controlunits, with moderate power (550-kw peak), wavelength 23 ern, and MovingTarget Indicator circuitry (see Richardson, 1972 for details). In 1973at least one and occasionally as many as four observers were in the field

52

each day from early September to early October, and less regularly atother times. In addition to observing hawks passing specific points, weused real-time radar observations and a radio-telephone to attempt todirect an observer to the inland concentrations of hawks. All radar datain all years were recorded on time-lapse film for later analysis.

Figure 1. Lines of echoes from hawks visible on the PPI display of theAASR-l radar at Toronto, 13:30 EST on 14 September, 1973. Almost all ofthe point echoes also represent birds. The photo is a ten-second timeexposure and shows a single revolution of the radar beam. The displayshows an area of radius 40 nautical miles (46 statute miles; 74 km),with range rings at ten-nautical-mile intervals. See Figure 2 for a mapof the radar coverage area.

Lines of echoes from hawks appeared in the Toronto and London areasboth inland and near the lakes on various days in September and October.As expected on the basis of the visual observations, the concentrationsalong the shoreline were most pronounced east of Toronto and Hawk Cliff.West of these two locations the hawks usually moved inland rather thanturning to the left to remain along the shorelines (Figures I and 2).

53

Figure 2. Map of southwestern Ontario showing typical locations of linesof hawk echoes within the London and Toronto radar coverage areas (40­nautical-mile radius) in autumn.

50 MI

80KM

In the London area the main flight line was often directly over theshore from 15 or more miles to the east of Hawk Cliff to Hawk Cliff itself.Near Toronto the main line east of the city was usually a few miles in­land rather than directly over the shoreline. On some days it was as faras eight or ten miles inland. Furthermore, in both the London and theToronto areas, lines of echoes were very frequent even farther inland.In the Toronto area, these lines were often visible as far as 35 nauticalmiles north-northwest of the radar site, or about 50 statute miles northof Lake Ontario. The echoes in these lines far to the north were usuallymoving to the southwest, whereas those closer to the lake were usuallymoving west-southwest.

The locations of the lines well inland from the north shores ofLakes Ontario and Erie were highly variable. Some lines remained inroughly the same location throughout a given day, but there was littletendency for lines to recur on different days at some points to the exclu­sion of others. Thus on some specific days a few lines would be widely

54

spaced at constant locations, but during the next hawk flight they mightbe widely spaced elsewhere in the area.

Furthermore, I frequently detected lines of echoes whose positionchanged continuously. For example, on numerous days, most of the ENE-WSWaligned lines of echoes moved laterally southwards. The track of anindividual bird was therefore somewhat south of WSW, since it moved west­southwest along the line and simultaneously south with the line. On otherdays most of the lines remained over the same locations from hour to hour,and on a few days the lines moved laterally to the north. The variablelocations of inland concentrations of hawks are no doubt in large measureresponsible for the infrequency of inland sightings in southern Ontario.Very few inland observation points become recognized as good places to seemigrating hawks. Even using real-time radar surveillance and a radio­telephone link between the radar and field observers, we found that it wasvery difficult to obtain visual sightings on the main inland concentrations.Nonetheless large numbers do move across the area many miles inland fromthe lakes.

Distribution of hawks vs. weather

One of the previously unanswered questions that can be addressed byusing the radar data is the effect of weather on flight paths and concen­tration areas. I examined the relationships of various weather variablesto three measurements of hawk behavior and distribution near Toronto:

(i) the relative numbers within a few miles of Lake Ontario andfarther to the north;

(ii) the degree and direction of lateral motion by the lines ofechoes; and

(iii) the mean flight direction over the ground of the echoesinvolved in the lines.

Each of three characteristics of hawk migration is likely to be re­lated to a variety of meteorological variables, and perhaps to other kindsof variables. A number of multivariate techniques were used in the analyses,and care was taken to scale variables appropriately and to use proceduressuch as analysis of residuals to verify the accuracy of the assumptions(Richardson, 1974). The Biomedical (BMD) package of computer programs wasused to perform the analyses (Dixon, 1973).

Stepwise multiple discriminant analysis was used to determine whetherthere was any evidence that any weather variable was related to the north­south distribution of hawks. While the results of the various analyseswere less than perfectlY consistent, there was strong evidence that therewere relatively more hawks near the lake with winds having a northerly com­ponent and more inland with winds having a southerly component.

This result was extended by stepwise multiple regression analysis ofthe lateral motion of the lines versus weather conditions. The regressionanalysis showed a strong tendency for the lateral motion to be correlatedwith the north-south wind component and to related variables such as pres­sure trend and temperatur~. The lines tended to move laterally southwardswith northerly winds, and to remain stationary or move northward withsoutherly winds.

The mean flight directions of individual echoes showed no obviousrelationship to any weather variables. However, they were related to the

55

lateral motion of the lines. When the lines moved south, the birds tendedto be directed further to the left of wsw than when the lines were sta­tionary or moved north. These results were again obtained using stepwisemultiple regression.

The results clearly show that hawks tended to concentrate near LakeOntario with northerly winds more than they did with other winds. Themethod by which this concentration occurred is still not clear, partlybecause we have no clear understanding of the nature of the lines oflaterally-moving echoes. However, even the relatively stationary lines ofechoes inland over Ontario do not seem to be strongly associated with speci­fic topographic features. This enhances the probability that atmosphericphenomena, or their interactions with topography, were responsible for theconcentrations inland. It is probable that there are various reasons forthe formation of lines of migrating hawks. None of the phenomena describedby John Haugh elsewhere in this volume (page 72) could alone account for allof the situations in which lines of hawks have been detected. Simultaneousvisual, long-range surveillance radar, and high-resolution surveillance ortracking radar data would be very useful in clarifying the situation. Inaddition, more precise data on atmospheric conditions are needed.

Volume of migration vs. weather

Another topic that was examined using the radar data was the day-to­day variation in numbers aloft. The analysis was based on the entire Torontoradar-coverage area, whereas visual observers around Toronto and Hawk Cliffhave based their conclusions on numbers seen along or near the shore. InSeptember, hawk movements were very significantly more frequent and denserin synoptic weather conditions involving north or east winds than in thoseinvolving south or west winds. In other words, there were many hawks aloftafter the passage of a cold front as a low moved away to the east, a highapproached or moved past to the north, or both. Large numbers were alsoaloft when there was a high pressure area nearby. These results are similarto those derived by subjective assessment of previous visual observationsalong the lakeshores.

However, more detailed analysis of the radar data showed that conclu­sions based on these visual observations alone are biased. MUltivariateanalysis of numbers versus 15 weather variables showed a strong relationshipbetween the numbers aloft, fair weather, and winds having an easterly com­ponent. The greater correlation with easterly than with northerly windswas very clear. This result was confirmed by a more sophisticated analysis,in which we first used factor analysis to identify three basic meteorologi­cal dimensions amongst the 15 original weather variables, and then relatedthe numbers aloft to these weather factors using discriminant analysis.

Comparison of radar and visual results

Hawk migration was detectable by both radar and visual techniques.Visual observations provided more detailed but intermittent and localizeddata about the identity of the birds, about flock sizes, and about numbersaloft. Each radar provided more continuous data for all parts of an areaabout 80 miles in diameter. The radars revealed previously unknown flightpaths, and permitted study of migration patterns even on days when onlysmall and visually inconspicuous flights were occurring. If a tra~ing

radar or a surveillance radar with high resolution had been available, moredetailed information could have been obtained about flight paths.

The broad-scale data obtained by radar permitted the first systematic

56

analysis of day-to-day variations in the geographic distribution and flightdirections of hawks migrating over Ontario. These results, together withanalyses of migration volume, showed that visual observations at knownshoreline concentration points provide a biased impression of flight pat­terns and of weather conditions optimal for migration across southern Ontarioas a whole.

Visual observations at Hawk Cliff, Ontario; Cedar Grove, Wisconsin;Derby Hill, New York, and of some species at Hawk Mountain show strong cor­relations between side winds rather than following winds, and numbers aloft(Mueller and Berger, 1961, 1967a, b; Haugh, 1972). In contrast, peak volumesof migration in species other than hawks are closely related to followingwindS. The radar study of hawks near Toronto suggests that the apparentrelationship of peak numbers to side (northwest and north) winds is anartifact of overly localized observations; maximal numbers do migrate acrosssouthern Ontario as a whole when winds are following (NE-E). It is quitepossible that similar phenomena occur near the other concentration pointsnoted above. Specifically, weather conditions that concentrate maximalnumbers of hawks at known lookouts may not produce the densest movementsacross the broad area around those lookouts. Radar data from those areaswould give a broader perspective and help to answer many of the outstandingquestions about the flight behavior of hawks.

ACKNOWLEDGEMENTS

Our radar studies were begun by the Associate Committee on BirdHazards to Aircraft of the National Research Council of Canada under thedirection of Dr. W. W. H. Gunn. More recently the studies in the Torontoarea were financed by the Toronto Area Airports Project, Canada Ministryof Transport, through a contract with LGL Ltd. M. McLaren assisted withdata compilation; R. A. Davis, T. Davis, J. O. L. Roberts, and G. Yakimade the field observations.

REFERENCES

Able, K. P. 1970. A radar study of the altitude of nocturnal passerinemigration. Bird-Banding 41: 282-290.

Alerstam, T. and S. Ulfstrand.diurnal bird migration inScandinavica 3: 99-139.

1972. Radar and field observations ofSouth Sweden, autumn 1971. Ornis

Bruderer, B., and P. Steidinger. 1972. Methods of quantitative andqualitative analysis of bird migrations with a tracking radar.In Animal Orientation and Navigation, ed. S. R. Galler et al.NASA SP-262. U.S. Govt. Printing Office, Washington.

Dixon, W. J. 1973. BMD: Biomedical computer programs. univ. CaliforniaPress, Berkeley, 773 pp.

Eastwood, E. 1967. Radar ornithology. Methuen, London, 278 pp.

Evans, P. R. and G. W. Lathbury. 1973. Raptor migration across the Straitsof Gibraltar. Ibis 115: 572-585.

Gauthreaux, S. A., Jr. 1970. Weather radar quantification of birdmigration. BioScience 20: 17-20.

57

L

Gehring, W. 1963. Radar und Feldbeobachtungen Uber den Verlauf desVogelzuges im Schweizerischen Mittelland: Der Tagzug im Herbst(1957-1961). Der Ornithologische Beobachter 60: 35-68.

Gunn, W. W. H. 1954. Hints for hawk watchers. Federation of OntarioNaturalists Bulletin 65: 16-19.

Haugh, J. R. 1972. A study of hawk migration in eastern North America.Search 2: 1-60.

Houghton, E. W. 1970. Spring migration at Gibraltar. Royal Radar Estab- I

lishment Memorandum 2593, Malvern, Worcs., U.K.I

-------- 1974. Highlights of the NATO-Gibraltar bird migration radar I

study. Privately distributed copy of paper presented to Bird Strike ICommittee Europe, Paris COnference, May 1973. Royal Radar Establish­ment, Malvern, Worcs., U.K. I

Mueller, H. C., and D. D. Berger. 1961. Weather and fall migration ofhawks at Cedar Grove, Wisconsin. Wilson Bulletin 73: 171-192.

-------- 1967a. Wind drift, leading lines, and diurnal migration.Wilson Bulletin 79:50-63.

Fall migration of Sharp-shinned Hawks.-------- 1967b.79: 397-415.

Wilson Bulletin i

I

Nisbet, I. C. T. 1963. Measurements with radar of the height of nocturna~

migration over Cape Cod, Massachusetts. Bird-Banding 34: 57-67. I

Temporal variations in the ability of individual Ibirds. National Research Council of Canadaon Bird Hazards to Aircraft Field Note

Richardson, W. J. 1972.radars in detectingAssociate Committee61: 1-68. i

I

-------- 1974. Multivariate approaches to forecasting day-to-day varia- Itions in the amount of bird migration. Proceedin s of the Conferencon Biological Aspects of the Bird/Aircraft Collision Problem, ClemsoUniv., S.C. pp. 309-329.

Solman, V. E. F. 1969. Photography in bird control for air safety.Journal of Biological Photographers Association 37: 150-155.

Williams, T. C., and H. Mix. 1973.data from radar Plan position44: 124-125.

An inexpensive apparatus for recordin~Indicator displays. Bird-Banding ,

I

Williams, T. C., J. Settel, P. O'Mahoney, and J. M. Williams. 1972a. Anornithological radar. American Birds 26: 555-557.

Williams, T. C., J. M. Williams, J. M. Teal and J. W. Kanwisher. 1972b. I

Tracking radar studies of bird migration. In Animal Orientation and!Navigation, ed. S. R. Galler et al. NASA SP-262. U.S. Govt. I

Printing Office, Washington.

58

DISCUSSION OF TELEMETRY AND RADAR

QUESTION: Have broadwings been seen crossing the Great Lakes?

RICHARDSON: No. Sometimes we see them going out as much as a mile andflying parallel to the lake shore, but only when the winds are very strongfrom the north. But as far as birds heading across the lakes, no.

PERSHING HOFS LUND , Capt. Perkins here has of course traveled the GreatLakes for many many years and kept observations. Maybe he has some comment,too, along this line.

J. P. PERKINS: After two million miles of traveling over the waters of theGreat Lakes and book after book of observations of birds, the buteos--I cancount my sightings on both hands, and I have five fingers on each hand.Accipters, yes. We saw lots of sharpshins in past years. The sharpshinsand kestrels have deteriorated. Maybe once or twice every three years wesee a peregrine now. Ospreys and marsh hawks are also frequently seencrossing the lakes. There is a good hawk flyway from the eastern end ofIsle Royale to Keweenaw Peninsula--mostly Ospreys, a few roughlegs, marshhawks, sharpshins, and kestrels. They do cross the lakes in quite largenumbers. I see them from my ship. I see them on board. I've photographedthem. I did photograph one peregrine. It was late in the evening--got soexcited, I forgot to allow for the poor quality of the evening light. Butthe buteos: on August 20 back in the late ~Os a pair of immature Red­shouldered Hawks came aboard, and it was a beautiful afternoon. They werecrossing in the Apostle Island region of Lake Superior [off Duluth]. Comenighttime, I captured these two hawks and kept them in my room for 21 days.Jack Hofslund banded them and let them go on Hawk Ridge, Minnesota. I livedwith them for 21 days. I have detailed observations on their behavior, andthe other crew members could tell anytime I was within range of these twohawks because they started a peculiar whistling noise--I could go on andon telling you about birds over the lakes ... And in answer to the radarman, I tell you, when we're using our radar aboard ship, we're not interestedin birds. The new .•• radars we have, and some of the older ones, I can pickup any object on the water that's of interest to the navigator; and you'vegot to realize it's a potential danger to your ship. So we make an allow­ance for a target on the water ahead. We can't prove that it is stationary,but we take evasive action, and then to go by and see a seagull sitting onthe water--I could wring his damn neck.

DAVID BIRD: [Asks if radar has been used in stUdying owl migration.]

RICHARDSON: Not that I know of. I don't know that anybody has reporteddetecting owls on radar. Radars have difficulty in picking up things at lowelevations. Even radars such as the ones that we were using still don'tpick up things very well if they are flying low-level or over areas wherethere is intense... ground return. And the more short-range, high-resolutionradars such as the marine ones, which would be essential or which would bedesirable in trying to look for something visually, or trying to identifyit, don't have the capacity to suppress the ground echoes.

QUESTION: [On the unit cost- of telemetry.]

TOM DUNSTAN: On the average, the way the market looks now, there are aboutfive or six different companies that do this sort of thing. To buy one trans­mitter would probably range from about $60 to $120, depending on which trans­mitter you want. The receivers will vary from around $300 to $600. Now,

59-- -- -- - - ---.---

Can you tell us anything about delivery time on transmitters?TOM CADE:

the one I had up here was about the $600 deluxe model, 12 channels, and soon. Three-channel models are around $300, and I think there is now onethat's corning out that's going to bring it down to around $150 to $300, whicHis going to be a good receiver and looks pretty economical. Antennaes: as :for the Yagi (a Japanese type of antenna that looks like a fish skeletonl-­that goes for $9.95. Double-Yagis are twice that. A l6-elernent Yagi forthe really long distance stuff •••would be about $19.95. The "push-up"--up Ito forty feet--would be $22. The vw [which Dunstan used in his work] I don't

m~. !"

DUNSTAN: People are behind, people are ahead. If you get an order in andare lucky, it can be within two or three weeks. What happens is that we getpeaks in the fall and peaks in the spring, and then it gets quite slow. Youhave to commit a person to a date, and you have to hold them to it--any ofthe companies.

RICHARDSON: I've done a lot of radar work in Nova Scotia, and the answer isno. We picked up some things that might or might not have been hawks •..

in

andwho canability

who don't knowin the countryoutgrowing the

of youpeoplethink,

QUESTION: Has anybody tracked hawks on radar, particularly broadwings,Nova Scotia?

CADE: The real problem, of course, for thosearen't using them, is that there are very fewmake these things properly. The demand is, Ito produce, right now.

I,!,

!,I

rThere are big movements of hawks down Digby Neck in the northwest corner of !Nova Scotia, and I have picked up what looks like a line of something on the Iradar, moving southwest there. Whether it was hawks or not, I don't know. !It was too late for broadwings.

CADE: There is a telemetry information packet available through the RaptorResearch Foundation.

MARK FULLER: The Raptor Research Foundation has a telemetry committee, and I!recently we put together a package that consists of reports from people a- Iround the country that have done, or are doing, or anticipate doing, telemet 'research on raptors. Now this covers everything from physiological telemet~

to long-range tracking. This little pamphlet runs to about 48 pages and telabout each person's project, what they want to do, where they got their equiment, the approximate costs. This is available for $1.00 a copy from ByronHarrell:

Raptor Research Foundation, Inc.,c/o Biology Department,University of South Dakota,Vermillion, S. D. 57069

60